The author says: "The most prestigious universities – the Ivy League, University of Chicago, Stanford University, the University of California system – tend to lie in the most expensive parts of the country. Even with full funding, it is nearly impossible to live in such costly cities without incurring debt, given that stipends tend to be $25,000 or less."
Such entitlement. I got into UC Berkeley (1995-2000) where I received $13K/year to live on from them (for TA'ing), and I made it work by having a roommate, being frugal, and getting some student loans (the minimum I needed), which I paid off within 2 years of graduation. It sucked, but I had to just be lean and deal with it, and reminded myself that I was a student.
Also, there is a naive and incorrect (in math at least) assumption in this article that the only difference between a graduate student from a top-10 university and a top-100 university is the name of the university where they graduated. There may be some truth to this for undergraduates, but for graduate students---who spend between 4 and 10 years busting their ass full-time on their research area---the situation is completely different.
The author thus seems to discount the possibility that the very best students who go to graduate school at Harvard/Berkeley/etc. are often be the best job candidates. Harvard/Berkeley/etc. gets the best potential grad students in the first place -- I was on the grad admissions committee in math at Harvard for years, and we would make offers to the top (or so) American students, and pretty much get most of them. These were extremely strong enthusiastic students who had invested often a decade of their lives into mathematics already; the education these students then got at grad school was by top researchers. In research, learning from the masters is often critical to becoming one, since the habits, tricks, ideas, intuitions, etc. are very hard to convey in writing.
Also, not everybody gets into graduate school at Harvard/Berkeley/etc. via "inherited wealth". At least I didn't -- I was an undergraduate at Northern Arizona University, and got into Berkeley entirely by being crazy in love with doing mathematics and also relatively good at it. The undergrad institution definitely matters in grad school admissions. However, being sufficiently good regularly gets around being from a non-top-10 university.
The article reminds me of this quote -- http://en.wikiquote.org/wiki/Paul_Halmos: "André Weil suggested that there is a logarithmic law at work: first-rate people attract other first-rate people, but second-rate people tend to hire third-raters, and third-rate people hire fifth-raters."